Corona Virus Analysis¶
By Anubhav Sharma
Importing datasets¶
In [1]:
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
In [2]:
df_confirmed=pd.read_csv('https://raw.githubusercontent.com/CSSEGISandData/COVID-19/master/csse_covid_19_data/csse_covid_19_time_series/time_series_covid19_confirmed_global.csv')
df_confirmed.head()
Out[2]:
In [3]:
df_recovered=pd.read_csv('https://raw.githubusercontent.com/CSSEGISandData/COVID-19/master/csse_covid_19_data/csse_covid_19_time_series/time_series_covid19_recovered_global.csv')
df_recovered.head()
Out[3]:
In [4]:
df_death=pd.read_csv('https://raw.githubusercontent.com/CSSEGISandData/COVID-19/master/csse_covid_19_data/csse_covid_19_time_series/time_series_covid19_deaths_global.csv')
df_death.head()
Out[4]:
In [5]:
df_cases=pd.read_csv('https://raw.githubusercontent.com/CSSEGISandData/COVID-19/web-data/data/cases_country.csv')
df_cases.head()
Out[5]:
In [6]:
print(df_confirmed.shape)
print(df_death.shape)
print(df_recovered.shape)
print(df_cases.shape)
Data Preprocessing¶
In [7]:
#All the first 3 datasets have same number of columns
df_confirmed.columns
Out[7]:
In [8]:
df_cases.isnull().sum()
Out[8]:
In [9]:
#let's see where we stand on Global level
df_global=df_cases.copy().drop({'Last_Update','Lat','Long_','Country_Region'},axis=1)
global_summary=pd.DataFrame(df_global.sum())
In [10]:
global_summary=global_summary.transpose()
global_summary
Out[10]:
In [11]:
global_summary=global_summary.drop({'People_Tested','People_Hospitalized','Incident_Rate','UID'},axis=1)
global_summary.style.format("{:,.0f}")
Out[11]:
In [12]:
#Now we will visualize Covid 19 Clusters in the world
import matplotlib.pyplot as plt
import matplotlib.cm as cm
import matplotlib.colors as colors
import folium
In [13]:
print(df_cases['Lat'].isnull().sum())
print(df_cases['Long_'].isnull().sum())
In [14]:
df_cases[df_cases['Lat'].isnull()]
Out[14]:
In [15]:
df_cases=df_cases.drop(df_cases.index[[68,116]])
df_cases['Lat'].isnull().sum()
Out[15]:
Visualizing covid 19 across the world¶
Most number of cases¶
We can see that most number of cases are in North and South America followed by India
In [16]:
#total confirmed cases around the world
world_map_confirmed=folium.Map(location=[20,0],tiles='cartodbdark_matter',zoom_start=2)
for lat,lng,confirmed in zip(df_cases['Lat'],df_cases['Long_'],df_cases['Confirmed']):
label='{}'.format(confirmed)
label=folium.Popup(label,parse_html=True)
folium.CircleMarker([lat,lng],
radius=confirmed/60000,
popup=label,
color='crimson',
fill_color='#DC143C ',
fill=True,
fill_opacity=0.7,
parse_html=False).add_to(world_map_confirmed)
world_map_confirmed
Out[16]:
Most Number of recoveries¶
Here we can see many countries with good recovry number but most number of recoveries are in USA
In [17]:
#Total Recovered Cases
world_map_recovered=folium.Map(location=[20,0],tiles='cartodbdark_matter',zoom_start=2)
for lat,lng,recovered in zip(df_cases['Lat'],df_cases['Long_'],df_cases['Recovered']):
label='{}'.format(recovered)
label=folium.Popup(label,parse_html=True)
folium.CircleMarker([lat,lng],
radius=recovered/50000,
popup=label,
color='#32CD32',
fill_color='#32CD32',
fill=True,
fill_opacity=0.7,
parse_html=False).add_to(world_map_recovered)
world_map_recovered
Out[17]:
Most number of deaths¶
so here we can see an interesting visual which tells us that some countries who are not even in top 5 countries having most number of cases have signiificant amount of deaths. India on the other hand have managed to control death rate.
In [18]:
#total number of deaths
world_map_death=folium.Map(location=[20,0],tiles='cartodbdark_matter',zoom_start=2)
for lat,lng,death in zip(df_cases['Lat'],df_cases['Long_'],df_cases['Deaths']):
label='{}'.format(death)
label=folium.Popup(label,parse_html=True)
if(death==0):
pass
else:
folium.CircleMarker([lat,lng],
radius=death/5000,
popup=label,
color='red',
fill_color='red',
fill=True,
fill_opacity=0.7,
parse_html=False).add_to(world_map_death)
world_map_death
Out[18]:
In [19]:
ts_confirmed=pd.DataFrame(df_confirmed.sum()).transpose()
ts_confirmed
Out[19]:
Visualizing daily increase , recoveries and deaths of covid 19 around the world¶
Here we will see that how have we done as a world in this pandemic
In [20]:
ts_confirmed=ts_confirmed.drop({'Lat','Long'},axis=1)
In [21]:
ts_confirmed=ts_confirmed.transpose()
ts_confirmed.head()
Out[21]:
In [22]:
ts_confirmed=ts_confirmed.reset_index()
In [23]:
ts_confirmed.head()
Out[23]:
In [24]:
ts_confirmed=ts_confirmed.rename(columns={'index':'dates',0:'num_cases'})
ts_confirmed.head()
Out[24]:
In [25]:
import plotly.express as px
import plotly.offline as pyo
import plotly.graph_objects as go
pyo.init_notebook_mode()
In [26]:
ts_recovered=pd.DataFrame(df_recovered.sum())
ts_recovered.head()
Out[26]:
In [27]:
ts_recovered=ts_recovered.drop({'Lat','Long'},axis=0)
ts_recovered=ts_recovered.reset_index()
ts_recovered=ts_recovered.rename(columns={'index':'dates',0:'recovered'})
ts_recovered.head()
Out[27]:
In [28]:
ts_death=pd.DataFrame(df_death.sum())
ts_death.head()
Out[28]:
In [29]:
ts_death=ts_death.drop({'Lat','Long'},axis=0)
ts_death=ts_death.reset_index()
ts_death=ts_death.rename(columns={'index':'dates',0:'deaths'})
ts_death.head()
Out[29]:
In [30]:
# Now we will create a dataframe for active
#series converts normal array into columns
ts_active=pd.Series(
data=np.array(
[x1-x2-x3 for (x1,x2,x3) in zip(ts_confirmed['num_cases'],ts_death['deaths'],ts_recovered['recovered'])]),index=ts_confirmed['dates'])
In [31]:
ts_active=ts_active.reset_index()
ts_active.head()
Out[31]:
In [32]:
ts_active=ts_active.rename(columns={0:'active'})
ts_active.head()
Out[32]:
In [33]:
# Now we will visualize number of cases,active cases,total recoveries and total deaths
trace0=go.Scatter(
x=ts_confirmed['dates'],
y=ts_confirmed['num_cases'],
mode='lines',
name='Total num of cases')
trace1=go.Scatter(
x=ts_recovered['dates'],
y=ts_recovered['recovered'],
mode='lines',
name='Total recoveries')
trace2=go.Scatter(
x=ts_death['dates'],
y=ts_death['deaths'],
mode='lines',
name='Total deaths')
trace3=go.Scatter(
x=ts_active['dates'],
y=ts_active['active'],
mode='lines',
name='Total active cases')
In [34]:
data=[trace0,trace1,trace2,trace3]
layout=go.Layout(title='Corona Virus scenario')
fig=go.Figure(data,layout)
fig.show()
In [35]:
fig = px.bar(ts_confirmed, x="dates", y='num_cases', title="daily confirmed cases")
fig.show()
Top 20 countrues with most number of cases¶
In [36]:
#Now we will see top 20 countries
df_top20=df_cases.drop({'Last_Update','Lat','Long_','People_Hospitalized','Mortality_Rate','UID','ISO3'},axis=1)
In [37]:
df_top20.sort_values(by='Confirmed',axis=0,ascending=False,inplace=True)
In [38]:
df_top20=df_top20.head(20)
In [39]:
df_top20.reset_index(drop=True,inplace=True)
In [40]:
df_top20=df_top20.drop({'People_Tested'},axis=1)
In [41]:
df_top20.style.bar(align='left',subset=['Confirmed', 'Deaths','Recovered','Active','Incident_Rate'], color='Red')
Out[41]:
Analysing the situation in India¶
Now we will analyse the situation of coronavirus in India
In [42]:
#Now we will Analyze India
df_india=pd.DataFrame(df_top20.loc[3,:])
In [43]:
df_india=df_india.transpose()
In [44]:
df_india.reset_index(drop=True,inplace=True)
df_india
Out[44]:
In [45]:
ts_confirmed_india=df_confirmed[df_confirmed['Country/Region'] == 'India']
ts_confirmed_india=ts_confirmed_india.drop({'Province/State','Lat','Long'},axis=1).reset_index(drop=True).sum()
ts_confirmed_india=ts_confirmed_india.to_frame()
ts_confirmed_india=ts_confirmed_india.drop(ts_confirmed_india.index[0])
ts_confirmed_india=ts_confirmed_india.reset_index()
ts_confirmed_india=ts_confirmed_india.rename(columns={'index':'dates',0:'confirmed'})
ts_confirmed_india.head()
Out[45]:
In [46]:
ts_recovered_india=df_recovered[df_recovered['Country/Region']=='India']
ts_recovered_india=ts_recovered_india.drop({'Province/State','Lat','Long'},axis=1).reset_index(drop=True).sum()
ts_recovered_india=ts_recovered_india.to_frame()
ts_recovered_india=ts_recovered_india.drop(ts_recovered_india.index[0])
ts_recovered_india=ts_recovered_india.reset_index()
ts_recovered_india=ts_recovered_india.rename(columns={'index':'dates',0:'recovered'})
ts_recovered_india.head()
Out[46]:
In [47]:
ts_death_india=df_death[df_death['Country/Region']=='India']
ts_death_india=ts_death_india.drop({'Province/State','Lat','Long'},axis=1).reset_index(drop=True).sum()
ts_death_india=ts_death_india.to_frame()
ts_death_india=ts_death_india.drop(ts_death_india.index[0])
ts_death_india=ts_death_india.reset_index()
ts_death_india=ts_death_india.rename(columns={'index':'dates',0:'deaths'})
ts_death_india.head()
Out[47]:
In [48]:
ts_active_india=pd.Series(
data=np.array(
[x1-x2-x3 for (x1,x2,x3) in zip (ts_confirmed_india['confirmed'],ts_death_india['deaths'],ts_recovered_india['recovered'])]),
index=ts_confirmed_india['dates']
)
ts_active_india=ts_active_india.to_frame()
ts_active_india=ts_active_india.reset_index()
ts_active_india=ts_active_india.rename(columns={'index':'dates',0:'active'})
ts_active_india.head()
Out[48]:
In [49]:
trace0=go.Scatter(
x=ts_confirmed_india['dates'],
y=ts_confirmed_india['confirmed'],
mode='lines',
name='Total num of cases')
trace1=go.Scatter(
x=ts_recovered_india['dates'],
y=ts_recovered_india['recovered'],
mode='lines',
name='Total recoveries')
trace2=go.Scatter(
x=ts_death_india['dates'],
y=ts_death_india['deaths'],
mode='lines',
name='Total deaths')
trace3=go.Scatter(
x=ts_active_india['dates'],
y=ts_active_india['active'],
mode='lines',
name='Total active cases')
Good rate of recovery¶
Even though cases are increasing significantly but we have managed to flatten active cases curve and we are having good number of recoveries while supressing the death curve
In [50]:
data=[trace0,trace1,trace2,trace3]
layout=go.Layout(title='Corona Virus scenario in India')
fig=go.Figure(data,layout)
fig.show()
In [51]:
fig = px.bar(ts_confirmed_india, x="dates", y='confirmed', title="Daily confirmed cases")
fig.show()
Analyising situation in various states across India¶
In [52]:
#Now I will Analyze India's data state wise
india_state=pd.read_csv('https://raw.githubusercontent.com/imdevskp/covid-19-india-data/master/state_level_latest.csv')
In [53]:
india_state.head()
Out[53]:
In [54]:
india_state.drop({'Last_Updated_Time','Migrated_Other','Delta_Confirmed','Delta_Recovered','Delta_Deaths','State_Notes'},
axis=1,inplace=True)
In [55]:
india_state.drop(india_state.index[[0,9]],inplace=True)
india_state=india_state.reset_index(drop=True)
india_state.style.bar(subset=['Confirmed','Recovered','Deaths','Active'],color='red')
Out[55]:
States with most number of cases¶
In [56]:
#lets create a bar chart of top 5 States comparing the number of confirmed cases
df_bar=india_state.loc[0:4,['State','Confirmed']]
In [57]:
df_bar.set_index('State',drop=True,inplace=True)
In [58]:
df_bar.plot(kind='bar',figsize=(10,6),rot=90,color='orange')
plt.xlabel('States')
plt.ylabel('Total Confirmed Cases')
plt.title('Top 5 States with Most number of Confirmed Cases')
plt.show()
States with most number of recoveries¶
Here we see that though Tamil Nadu has more number of cases delhi is leading in number of recoveries
In [59]:
#state with max recoveries
df_bar_recovered=india_state.loc[0:4,['State','Recovered']]
df_bar_recovered.set_index('State',drop=True,inplace=True)
df_bar_recovered.sort_values('Recovered',axis=0,ascending=False,inplace=True)
In [60]:
df_bar_recovered.plot(kind='bar',figsize=(10,6),rot=90,color='green')
plt.xlabel('states')
plt.ylabel('Number of recoveries')
plt.title('States with Most recoveries')
plt.show()
States with most number of deaths¶
Here we see that Tamil Nadu has managed to contain deaths even though having second most number of cases
In [61]:
#State with Max deaths
df_bar_death=india_state.loc[0:4,['State','Deaths']]
df_bar_death.set_index('State',drop=True,inplace=True)
df_bar_death.sort_values('Deaths',axis=0,ascending=False,inplace=True)
In [62]:
df_bar_death.plot(kind='bar',figsize=(10,6),rot=90,color='red')
plt.xlabel('states')
plt.ylabel('Number of deaths')
plt.title('States with Most Deaths')
plt.show()
#Tamil Nadu despite having Second largest cases is on 4th number in deaths
Here we wil analyse the state with most number of cases per million¶
In [63]:
import xlsxwriter
path=r'C:\Users\Anubhav\Untitled Folder\\Indian States Population and Area.xlsx'
df = pd.ExcelFile(path)
In [64]:
df.sheet_names
Out[64]:
In [65]:
df_pop = df.parse("Sheet1")
In [66]:
df_pop.head()
#I will use this data since we can't Rely on data of 2011 Census and this gives pretty basic Idea
Out[66]:
In [67]:
df_state_pop=pd.merge(india_state,df_pop,on='State')
In [68]:
df_state_pop.rename(columns={'Aadhaar assigned as of 2019':'Population'},inplace=True)
In [69]:
df_state_pop.head()
Out[69]:
In [70]:
df_state_pop['Cases/million']=(df_state_pop['Confirmed']/df_state_pop['Population'])*1000000
df_state_pop.sort_values('Cases/million',axis=0,ascending=False,inplace=True)
df_state_pop.reset_index(drop=True,inplace=True)
df_state_pop.drop({'Area (per sq km)'},axis=1,inplace=True)
df_state_pop.head()
Out[70]:
In [71]:
#We will visualize 5 states with most cases/million
df_bar_pop=df_state_pop.loc[0:4,['State','Cases/million']]
df_bar_pop.set_index('State',inplace=True)
Here we see an interesting thing that compared to its population Delhi has a lot more cases as compared to Maharasthra¶
In [72]:
df_bar_pop.plot(kind='bar',figsize=(10,6))
plt.xlabel('States')
plt.ylabel('Cases/million')
plt.title('States with most cases/million')
plt.show()
In [73]:
import seaborn as sns
plt.figure(figsize = (12,8))
sns.heatmap(df_state_pop.corr(), annot=True)
plt.show()
In [74]:
covid_daily=pd.read_csv('https://raw.githubusercontent.com/imdevskp/covid-19-india-data/master/state_level_daily.csv')
In [75]:
covid_daily.head()
Out[75]:
We will here Analyse which state among top 6 has managed to flatten the curve¶
In [76]:
#Now we will see top 6 states and which were able to flatten the curve
covid_maha=covid_daily[covid_daily['State']=='MH']
covid_maha=covid_maha.drop({'Deceased','Recovered','Unnamed: 0'},axis=1)
covid_maha['Date'] =pd.to_datetime(covid_maha.Date)
covid_maha.sort_values(by='Date',ascending=True,inplace=True)
covid_maha.reset_index(drop=True,inplace=True)
covid_maha.head()
Out[76]:
In [77]:
covid_del=covid_daily[covid_daily['State']=='DL']
covid_del=covid_del.drop({'Deceased','Recovered','Unnamed: 0'},axis=1)
covid_del['Date'] =pd.to_datetime(covid_del.Date)
covid_del.sort_values(by='Date',ascending=True,inplace=True)
covid_del.reset_index(drop=True,inplace=True)
covid_del.head()
Out[77]:
In [78]:
covid_guj=covid_daily[covid_daily['State']=='GJ']
covid_guj=covid_guj.drop({'Deceased','Recovered','Unnamed: 0'},axis=1)
covid_guj['Date'] =pd.to_datetime(covid_guj.Date)
covid_guj.sort_values(by='Date',ascending=True,inplace=True)
covid_guj.reset_index(drop=True,inplace=True)
covid_guj.head()
Out[78]:
In [79]:
covid_tn=covid_daily[covid_daily['State']=='TN']
covid_tn=covid_tn.drop({'Deceased','Recovered','Unnamed: 0'},axis=1)
covid_tn['Date'] =pd.to_datetime(covid_tn.Date)
covid_tn.sort_values(by='Date',ascending=True,inplace=True)
covid_tn.reset_index(drop=True,inplace=True)
covid_tn.head()
Out[79]:
In [80]:
covid_up=covid_daily[covid_daily['State']=='UP']
covid_up=covid_up.drop({'Deceased','Recovered','Unnamed: 0'},axis=1)
covid_up['Date'] =pd.to_datetime(covid_up.Date)
covid_up.sort_values(by='Date',ascending=True,inplace=True)
covid_up.reset_index(drop=True,inplace=True)
covid_up.head()
Out[80]:
In [81]:
covid_rj=covid_daily[covid_daily['State']=='RJ']
covid_rj=covid_rj.drop({'Deceased','Recovered','Unnamed: 0'},axis=1)
covid_rj['Date'] =pd.to_datetime(covid_rj.Date)
covid_rj.sort_values(by='Date',ascending=True,inplace=True)
covid_rj.reset_index(drop=True,inplace=True)
covid_rj.head()
Out[81]:
In [82]:
trace0=go.Scatter(
x=covid_maha['Date'],
y=covid_maha['Confirmed'],
mode='lines',
name='Cases in Maharashtra')
trace1=go.Scatter(
x=covid_del['Date'],
y=covid_del['Confirmed'],
mode='lines',
name='Cases in Delhi')
trace2=go.Scatter(
x=covid_tn['Date'],
y=covid_tn['Confirmed'],
mode='lines',
name='Cases in Tamil Nadu')
trace3=go.Scatter(
x=covid_up['Date'],
y=covid_up['Confirmed'],
mode='lines',
name='Cases in Uttar Pradesh')
trace4=go.Scatter(
x=covid_rj['Date'],
y=covid_rj['Confirmed'],
mode='lines',
name='Cases in Rajasthan')
trace5=go.Scatter(
x=covid_guj['Date'],
y=covid_guj['Confirmed'],
mode='lines',
name='Cases in Gujrat')
Here we see that Rajasthan , Gujrat and Uttar Pradesh have almost managed to flatten their curve despite having a lot cases in beginning of pandemic¶
In [83]:
data=[trace0,trace1,trace2,trace3,trace4,trace5]
layout=go.Layout(title='Which States are flattening Curve')
fig=go.Figure(data,layout)
fig.show()
Analysing Medical scenario in various states¶
In [84]:
# Now we will see which States are ready to face pandemic
df_health=pd.read_csv(r'C:\Users\Anubhav\Untitled Folder\datasets_557629_1259476_HospitalBedsIndia.csv')
df_health.head()
Out[84]:
In [85]:
df_health.columns
Out[85]:
In [86]:
df_health=df_health.drop({'Sno','NumPrimaryHealthCenters_HMIS','NumCommunityHealthCenters_HMIS','NumSubDistrictHospitals_HMIS',
'NumDistrictHospitals_HMIS','TotalPublicHealthFacilities_HMIS','NumRuralHospitals_NHP18',
'NumUrbanHospitals_NHP18'},axis=1)
df_health.head()
Out[86]:
In [87]:
df_health['Total Beds'] = df_health.loc[:,'NumPublicBeds_HMIS'].add(df_health.loc[:,'NumRuralBeds_NHP18']).add(df_health.loc[:,'NumUrbanBeds_NHP18'])
In [88]:
df_health.drop({'NumPublicBeds_HMIS','NumRuralBeds_NHP18','NumUrbanBeds_NHP18'},axis=1,inplace=True)
In [89]:
df_health=df_health.rename(columns={'State/UT':'State'})
df_health.head()
Out[89]:
In [90]:
df_state_health=pd.merge(df_state_pop,df_health,on='State')
df_state_health.head()
Out[90]:
In [91]:
#We can see that our top 5 affected states are not even in top 10 of beds/million index
df_state_health['beds/million']=(df_state_health['Total Beds']/df_state_health['Population'])*1000000
df_state_health.sort_values('beds/million',axis=0,ascending=False,inplace=True)
df_state_health.reset_index(drop=True,inplace=True)
df_state_health
Out[91]:
In [92]:
df_bar_beds=df_state_health.loc[0:5,['State','beds/million']]
df_bar_beds.set_index('State',drop=True,inplace=True)
df_bar_beds.head()
Out[92]:
Not even one state from top 5 states having most number of cases is in top 10 states of most beds/ million which tells us that we were not ready to face this pandemic¶
In [93]:
df_bar_beds.plot(kind='bar',figsize=(10,6))
plt.xlabel('States')
plt.ylabel('beds/million')
plt.title('States with most beds per million')
plt.show()
In [ ]: